WO2022261961A1

WO2022261961A1 - Method for forwarding message, and switch and network interface card

Info

Publication number: WO2022261961A1
Application number: PCT/CN2021/101018
Authority: WO
Inventors: 郑盛巍; 刘亚社; 吕超; 周轶刚
Original assignee: 华为技术有限公司
Priority date: 2021-06-18
Filing date: 2021-06-18
Publication date: 2022-12-22
Also published as: CN117480744A

Abstract

The embodiments of the present application relate to a method for forwarding a message, and a switch and a network interface card. The method for forwarding a message comprises: a switch receiving a message from a network interface card; and the switch forwarding the message according to a configurable parameter in a destination media access control domain in a frame format of the message. In this way, the accelerated forwarding of a message can be realized, so as to reduce a delay during a forwarding process. In addition, the complexity during the forwarding process can be significantly reduced, thereby improving the bandwidth efficiency, and reducing the delay.

Description

Method for forwarding message, switch and network interface card

technical field

Embodiments of the present disclosure relate to electronic devices. More specifically, it relates to a method for forwarding packets, a switch and a network interface card.

Background technique

Based on standard Ethernet frames and standard Internet Protocol (Internet Protocol, IP) messages, the first Transmission Control Protocol/Internet Protocol (Transmission Control Protocol/Internet Protocol, TCP/IP) is currently in the data center network (Data Center Network, DCN) adopts the most extensive protocol type, followed by the Remote Direct Memory Access (RDMA) protocol carried on standard IP and standard Ethernet has also increased. In a switch fabric-based DCN, most of the various fabric protocols based on the packet-by-packet load balancing technology used to increase bandwidth are basically carried by standard Ethernet frame formats.

Contents of the invention

Embodiments according to the present disclosure relate to a method for forwarding packets, a switch and a network interface card.

In the first aspect of the embodiments of the present disclosure, a method for forwarding packets is provided. The method includes the switch receiving the message from the network interface card. The method also includes the switch forwarding the message according to the configurable parameters in the destination media access control field in the frame format of the message, wherein the configurable parameters are associated with the message type of the message, and the message type includes The first type of packet in the Ethernet frame format and the second type of packet in the standard Ethernet frame format.

According to the method of the first aspect of the present disclosure, the switch can determine the type of the message according to the parameter at a specific position in the message frame format and forward the message based on the determined message type. Therefore, the switch can implement different processing processes for different types of messages, for example, a simplified forwarding process can be used for simplified Ethernet messages, and other forwarding methods can be used for regular messages. In this way, accelerated forwarding of packets can be realized to reduce the time delay in the forwarding process.

In addition, by setting parameters dedicated to a certain message type in the message frame structure, a simplified mode of the Ethernet frame format can be realized without changing the original frame structure. Since the definition of the simplified Ethernet frame format is more concise than that of the standard Ethernet, the link bandwidth can be improved.

In some implementation manners, the switch may compare the configurable parameter with the predetermined parameter of the destination MAC field in the reference frame format of the first type packet. If it is determined that the configurable parameter matches the predetermined parameter, the received message is determined as the first type message. The switch may determine an output port number associated with the destination address of the message and forward the message based on the output port number.

In some implementation manners, the switch may compare the configurable parameter with the predetermined parameter of the destination MAC field in the reference frame format of the first type packet. If it is determined that the configurable parameter does not match the predetermined parameter, then determine the received message as the second type of message. The switch may determine whether the message is allowed to be forwarded to the target device of the message by performing an inbound processing operation on the message. If it is determined that the message is forwarded to the destination device of the message via the switch, the switch may determine an output port number associated with the destination address of the message and forward the message based on the output port number. message.

The switch can determine the processing process for the message according to different message types, so that the steps in the message forwarding process can be correspondingly reduced, thereby reducing the possible delay in the forwarding process.

In some implementation manners, the switch obtains the destination address from the frame format of the packet. The target address includes one of the following: the address of the target device to receive the message, or the address of a target network segment related to the target device. The switch determines the output port number by performing an exact match operation or a linear table lookup operation based on the target address.

In some implementations, the address of the target network segment includes a configurable bitmask of predetermined length.

In some implementation manners, the packet may be processed in the inbound direction by using a tri-state content addressable memory matching operation.

In a second aspect of the embodiments of the present disclosure, a method for forwarding packets is provided. The method includes the network interface card determining a configurable parameter in a destination media access control field in a frame format of a message, wherein the configurable parameter is associated with a message type of the message, and the message type includes A first type of packet with a simplified Ethernet frame format and a second type of packet with a standard Ethernet frame format. The method also includes the network interface card generating the message based on the configurable parameter and sending the message to the switch.

According to the method of the second aspect of the present disclosure, the network interface card can implement a simplified Ethernet message without changing the original frame format by configuring the parameters of the specific field in the frame format of the message. Therefore, the packets using the simplified Ethernet frame format and the packets using the standard Ethernet frame format can not only intercommunicate, but also coexist. This simplified Ethernet message can significantly reduce the complexity of the forwarding process, improve bandwidth efficiency and reduce delay.

In some embodiments, the network interface card determines a type of data packet to be encapsulated into the packet and determines a configurable parameter based on the type of data packet.

In some embodiments, the network interface card obtains a set of predetermined parameters of the destination media access control field in the reference frame format associated with the first type of message and determines the configurable parameter from the set of predetermined parameters .

In a third aspect of the embodiments of the present disclosure, a switch is provided. The switch may include a processor and a memory coupled to the processor. The memory stores instructions that need to be executed, and when the instructions are executed by the processor, the switch executes any method according to the first aspect and its implementation manners.

In a fourth aspect of the embodiments of the present disclosure, a network interface card is provided. The network interface card may include a processor and a memory coupled with the processor. The memory stores instructions that need to be executed, and when the instructions are executed by the processor, the network interface card executes any method according to the second aspect and the implementation manners thereof.

In a fifth aspect of the present disclosure, a computer readable storage medium is provided. The computer-readable storage medium stores instructions, which, when executed by the electronic device, cause the electronic device to execute any method of the first aspect and its implementation manners or any method of the second aspect and its implementation manners.

In a sixth aspect of the present disclosure, a computer program product is provided. The computer program product includes instructions, which, when executed by the electronic device, cause the electronic device to perform any method of the first aspect and its implementations or any method of the second aspect and its implementations.

It should be understood that what is described in the Summary of the Invention is not intended to limit the key or important features of the embodiments of the application, nor is it intended to limit the scope of the application. Other features of the present application will become easily understood from the following description.

Description of drawings

The above and other features, advantages and aspects of the various embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. In the drawings, identical or similar reference numerals denote identical or similar elements, wherein:

Figure 1 shows a schematic diagram of an example environment in which an embodiment of the present application may be implemented;

FIG. 2 shows a signaling interaction diagram of an exemplary message forwarding process according to an embodiment of the present disclosure;

3A to 3C show schematic diagrams of example message types according to an embodiment of the present disclosure;

FIG. 4 shows a schematic diagram of a message forwarding process according to an embodiment of the present disclosure;

FIG. 5 shows a schematic diagram of a message forwarding process according to an embodiment of the present disclosure;

FIG. 6 shows a flowchart of a method for packet forwarding according to an embodiment of the present disclosure;

FIG. 7 shows a flowchart of a method for packet forwarding according to an embodiment of the present disclosure;

FIG. 8 shows a block diagram of a communication device according to an example embodiment of the present disclosure; and

FIG. 9 shows a block diagram of a communication device according to an example embodiment of the present disclosure; and

detailed description

Embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present application are shown in the drawings, it should be understood that the application may be embodied in various forms and should not be construed as limited to the embodiments set forth herein; A more thorough and complete understanding of the application. It should be understood that the drawings and implementations of the present application are for exemplary purposes only, and are not intended to limit the protection scope of the present application.

In the description of the embodiments of the present application, the term "comprising" and its similar expressions should be interpreted as an open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc. may refer to different or the same object. Other definitions, both express and implied, may also be included below.

In the current DCN, a typical network device may include a network interface card (Network Interface Card, NIC), such as a smart network card (Smart Network Interface Card, SNIC) and a data processing unit (Data Processing Unit, DPU). NICs can be regarded as end-side devices of various networks. Typical network equipment may also include switches. In DCN, the main types of frame formats mainly include various packets carried on the basis of standard Ethernet frames and standard IP packets, and standard Ethernet frames based on Switch Fabricized interfaces.

Among standard Ethernet frames and standard IP packets, TCP/IP is currently the most widely used protocol type in DCN, and secondly, the RDMA protocol carried on standard IP and standard Ethernet has also increased. In a switch-fabricized DCN, most of the various fabric protocols based on the packet-by-packet load balancing technology used to increase bandwidth are basically carried by the standard Ethernet frame format.

Since the standard Ethernet has established a strong ecological foundation in the WAN, the standard Ethernet frame format is still widely used in the DCN. In the Open System Interconnection Reference Model (OSI) standard seven-layer model, the first layer protocol (Layer 1, L1) in the standard seven-layer model, and the second layer protocol (Layer 2, L2) in the standard seven-layer model 1. The third layer protocol (Layer 3, L3) in the standard seven-layer model and the fourth layer protocol (Layer 4, L4) in the standard seven-layer model are most closely related to the DCN network. A typical configuration is that L2 adopts standard Ethernet frame format, L3 adopts standard Internet Protocol Version 4 (Internet Protocol Version 4, IPv4) or Internet Protocol Version 6 (Internet Protocol Version 6, IPv6), and L4 adopts TCP and User Datagram Protocol (User Datagram Protocol) Protocol, UDP). On the basis of the UDP protocol, it can also include a network protocol (RDMA over Converged Ethernet Version 2, RoCEv2) and an extended virtual local area network (Visual extend LAN, VxLAN) protocol that allows the use of RDMA over Ethernet.

In the frame format of the Switch Fabricized interface based on standard Ethernet frames, standard Ethernet frames are generally used at L2, followed by the Switch Fabricized private interfaces defined by each operator. The so-called DCN Switch Fabric technology refers to a type of technology in which each network operator applies the Switch Fabric technology that previously existed in a single network device to the entire DCN. Through the use of Switch Fabric technology in DCN, the bandwidth efficiency is greatly improved, so various Switch Fabric technologies are more and more widely used in DCN.

In addition, there is also a commonly used network frame format in the DCN, that is, a commercial supercomputer network protocol (Infiniband, IB) frame format. The IB network has low latency and good performance, making it the first choice for supercomputer (High Performance Computer, HPC) networks. However, because the scale of the IB network is generally small, IB is generally only used in HPC or storage internal networks. Another problem with the IB network is that the IB network cannot directly communicate with the DCN of the Ethernet, which makes the operation and maintenance of the cloud network using IB more difficult.

In the current DCN, standard Ethernet and IP packet network protocols, such as VxLAN, are used. There are problems such as large forwarding delay, large packet format overhead, and many redundant bytes. For example, although the standard Ethernet frame format of L2 exists, it becomes a redundant overhead field because it does not participate in the forwarding process, wasting network bandwidth. In addition, although the standard Ethernet frame format of L2 does not participate in forwarding, it participates in the parsing action (Parser) and matching action (Matching) of control packets, which introduces additional delay overhead.

Some network operators have proposed a packet header without an Ethernet frame format. In this scheme, the Ethernet packet header is removed from L2, and the forwarding and switching process is based on the L3 packet header. Compared with the standard Ethernet frame format, this solution can improve some bandwidth efficiency and increase some performance income. However, this scheme is not compatible with the existing standard Ethernet frame format. In actual use, existing software, management protocols, operation and maintenance protocols, routing protocols, etc. are not compatible, or need to be modified before they can run. The system overhead caused by this is huge. In addition, this solution does not bring corresponding improvement to the most important performance index of DCN, that is, the delay overhead in the process of forwarding and switching. Although IP packets without Ethernet frames are not processed at L2, control, management, and protocol packets are mixed together with service packets that require low latency.

The IB network uses its own L2 frame format and IPv6 for L3 forwarding, which can bring better network performance and lower delay. However, the IB network is small in scale, which is far inferior to the large-scale networking capability of Ethernet DCN. In addition, the IB network is not compatible with Ethernet. The intercommunication between IB network and Ethernet needs to be realized through a specific gateway. The existence of a dedicated protocol gateway not only affects the intercommunication performance, but also brings damage to the overall standardization of the DCN, making the operation and maintenance of the cloud network (that is, cloud DCN) troublesome, and the post-operation cost of the DCN is also increased.

Therefore, embodiments of the present disclosure provide a method for forwarding packets. The method includes the switch receiving the message from the network interface card. The method further includes the switch determining the message type of the message based on a configurable parameter in the destination media access control field in the frame format of the message. The method also includes the switch forwarding the message based on the determined message type. Thus, the switch can implement different processing processes for different types of messages, for example, a simplified forwarding process can be used for simplified Ethernet messages, and other forwarding methods can be used for regular messages. In this way, accelerated forwarding of packets can be realized to reduce the time delay in the forwarding process.

Figure 1 shows a schematic diagram of an example environment in which one embodiment of the present application may be implemented. As shown in FIG. 1, environment 100 may include NICs 110-1, 110-2, and 110-3. NICs 110-1, 110-2 and 110-3 may also be collectively referred to as NIC 110 hereinafter. NICs 110-1, 110-2, and 110-3 may be hosted on servers 101-1, 101-2, and 101-3, respectively. It should be understood that NIC 110 may be implemented in a variety of forms. For example, NIC 110 may be implemented including, but not limited to, a SNIC DPU, an intelligent top-of-rack (ToR) switch, and an intelligent edge switch, among others.

Environment 100 may also include switches 120-1, 120-2, 120-3, 120-4, and 120-5. The switches 120 - 1 , 120 - 2 , 120 - 3 , 120 - 4 and 120 - 5 may also be collectively referred to as switches 120 hereinafter. In environment 100, switches 120-1, 120-2, 120-3, 120-4, and 120-5 may be considered first-level switches (also referred to as ToR switches). As shown in FIG. 1, NICs 110-1, 110-2, and 110-3 can perform data transmission with switches 120-1, 120-2, and 120-3, respectively. The switches 120-1, 120-2 and 120-3 can respectively perform data transmission with the switches 120-4 and 120-5. It should be understood that switch 120 may be implemented in a variety of forms. For example, the switch 120 may be implemented to include but not limited to a ToR switch, a leaf switch (leaf switch), a spine switch (spine switch), and a core switch (core switch).

It should be understood that the NICs and switches included in environment 100 are not limited to those shown in FIG. 1 . Environment 100 may include any number of NICs and switches. Environment 100 is for exemplary purposes only and does not imply any limitation as to the scope of the present disclosure. Embodiments of the present disclosure may also be embodied in other network environments or architectures. In addition, it should also be understood that the environment 100 may also include other elements or entities for realizing communication connection, data transmission and other purposes. In order to simplify the description, these elements or entities are not shown in FIG. 1 , but it does not mean that the embodiments of the present disclosure do not have them.

The basic idea and principle of the present disclosure can be explained in detail through FIGS. 2 to 5 . FIG. 2 shows a signaling diagram 200 of an exemplary message forwarding process according to an embodiment of the present disclosure. In the following, the information interaction between the NIC 110 and the switch 120 during the message forwarding process according to the embodiment of the present disclosure will be described in detail first with reference to FIG. 2 . For clarity, the message forwarding process will be described in conjunction with the environment 100 shown in FIG. 1 .

As shown in FIG. 2, the NIC 110 may determine the configurable parameters in the destination MAC field in the frame format of the message and generate (205) the message according to the configurable parameters. The configurable parameter may be associated with the message type of the message. The packet type may include a first type packet in a simplified Ethernet frame format. The packet type may also include a second type packet in a standard Ethernet frame format.

In some embodiments, NIC 110 may determine what type of message to generate based on the type of traffic associated with the message to be generated. For example, if the service corresponding to the message to be generated is a service requiring low delay, the NIC 110 may determine that the message to be generated is the first type of message. And if the service corresponding to the message to be generated is a control class, a protocol class, an operation and maintenance class, or a management service, then the NIC 110 may determine that the message to be generated is a second type of message.

In some embodiments, if the NIC 110 is to generate a packet in a simplified Ethernet frame format, the NIC 110 can configure the corresponding simplified Ethernet frame format in the destination media access control (Destination MAC address, DMAC) domain in the frame format of the packet. Configurable parameters for the Ethernet frame format. For example, NIC 110 may obtain a set of predetermined parameters of the DMAC field in the reference frame format associated with the first type of message. These predetermined parameters may be preconfigured to NIC 110. The NIC may select parameters to be configured in the DMAC field in the frame format of the message from a set of acquired predetermined parameters.

FIG. 3A and FIG. 3B respectively show schematic diagrams of example packet types according to an embodiment of the present disclosure. Figure 3A shows an example of a standard Ethernet frame format, while Figure 3B shows an example of a simplified Ethernet frame format.

As shown in FIG. 3A , a standard Ethernet frame format may include an Ethernet header 310 . The Ethernet header 310 may include a DMAC field 311 , a source media access control (SMAC) field 312 , a VLAN field 313 , and an Ethernet frame checksum (FCS) field 314 and the like. The DMAC domain field 311 can include information about the destination address that the message is forwarded, the SMAC domain field 312 can include information about the source address of the message, and the FCS domain field can include information for checking the Ethernet frame for errors in the forwarding process. verification information.

As shown in FIG. 3B , the Ethernet header 320 of the simplified Ethernet frame format may include a DMAC domain field 321 , in which the predetermined DMAC domain parameters for the first type of message as described above are configured. Compared with the standard Ethernet frame format, one possible form of the simplified Ethernet frame format is that the Ethernet header in the simplified Ethernet frame format may not contain any L2 fields except the FCS field 322 . Furthermore, in another possible form of the simplified Ethernet frame format, the Ethernet header in the simplified Ethernet frame format may not contain any L2 fields except the FCS domain field 322 and a VLAN field (not shown).

Although the simplified Ethernet frame format simplifies some fields compared with the standard Ethernet frame format, it is still in the framework of the traditional Ethernet header in essence. Therefore, the simplified Ethernet frame format and the standard Ethernet frame format Text can not only communicate with each other, but also coexist. Therefore, through the specific definition of DMAC, it is possible to support the simplified Ethernet frame format of the message while being compatible with the existing rich ecology of the standard Ethernet message. Compatibility with the standard Ethernet frame format for sending and receiving means that it is compatible with a wide range of vibrant Ethernet ecosystems.

It can be seen that the NIC can realize a simplified Ethernet message without changing the original frame format by configuring the parameters of the specific field in the frame format of the message. This simplified Ethernet message can significantly reduce the complexity of the forwarding process, improve bandwidth efficiency and reduce delay.

In DCN, the current mainstream load sharing technology is a flow-by-flow load sharing method represented by Equal-cost multi-path (ECMP). ECMP involves a flow-level load balancing strategy that can reduce bursts and congestion on the network. However, the bandwidth efficiency of this load sharing method is low, so the Packet-based Dynamic Load Balance method will gradually become the mainstream of load sharing methods in DCN. To implement packet-by-packet load balancing, switch fabric can be introduced into the first type of packets.

In some embodiments, the Ethernet header 330 in the Ethernet frame format of the first type message generated by the NIC 110 may also include fields associated with Switch Fabric. FIG. 3C shows a schematic diagram of example message types according to an embodiment of the disclosure. As shown in FIG. 3C , the Ethernet header 330 may include a Switch Fabric header 332 in addition to the DMAC domain field 331 and the FCS field. The Switch Fabric head 332 can implement packet-by-packet load sharing technology, which not only has low byte overhead, high bandwidth efficiency, but also fast processing and low time delay. The Switch Fabric interface based on packet-by-packet load sharing technology can include source routing technology that carries multiple hop-by-hop switching port numbers and switching network technology that performs a linear table search for the destination port number at each hop.

Referring again to FIG. 2, after generating the message, NIC 110 sends (210) the generated message to switch 120.

After receiving a message from NIC 110, switch 120 may determine (215) how to forward the received message. The switch 120 may determine the forwarding mode of the message according to the configurable parameters in the DMAC domain in the frame format of the received message.

In some embodiments, for example, if the switch 120 can determine whether the configurable parameters in the DMAC field match the predetermined parameters in the DMAC field in the reference frame format of the first type message. If it is determined that the configurable parameters in the DMAC domain of the frame format of the received message match the predetermined parameters, the switch 120 may determine that the received message belongs to the first type of message. If it is determined that the configurable parameters in the DMAC domain of the frame format of the received message do not match the predetermined parameters, the switch 120 may determine that the received message belongs to the second type of message.

The switch 120 can forward packets according to different packet types. For the first type of message, that is, the message with a simplified Ethernet frame format, the switch 120 can skip the intermediate process of processing the message and directly forward the message, thereby obtaining excellent ultra-low delay characteristics. The intermediate process may include, for example, packet parsing, control management protocol packet matching, local packet matching, and the like. For the second type of message, that is, a message with a standard Ethernet frame format, the switch 120 can perform conventional Ethernet message forwarding processing by using channels such as message analysis.

As described above, the message type depends on the service type involved in the message. The first type of packets may involve packets associated with low-latency services, while the second type of packets may involve control packets, protocol packets, operation and maintenance Encapsulated ordinary service packets.

Fig. 4 shows an example of determining a packet forwarding manner according to an embodiment of the present disclosure. As shown in FIG. 4, after receiving the message 401, the switch will determine the message type of the message. If it is determined that the message is the first type message, then the message is directly forwarded or exchanged (430). And if switch determines that this message is the second type of message, then this message at first carries out message intermediate processing processes such as message parsing (410) and control message and local message matching (420), and then through intermediate processing The message will be forwarded and exchanged (430).

It can be seen that, compared with the second type of message, the first type of message can be forwarded faster, which can significantly reduce the delay caused by the intermediate process of the message. In addition, since the simplified Ethernet frame format of the first-type message is defined more concisely than the standard Ethernet of the second-type message, link bandwidth can be improved.

After determining the forwarding mode of the message, the switch 120 can forward or exchange the message received from the NIC 110. As mentioned above, for the first type of message, since the Ethernet header of the frame format of the first type of message includes the DMAC domain field with specific parameters, and no other L2 field is included in the Ethernet header except the FCS field. field, so the switch 120 can adopt a simplified forwarding process. As for the second type of message, although it has a standard Ethernet frame format, a relatively simplified forwarding process is also proposed in this application.

In some embodiments, if the switch 120 determines that the packet to be forwarded belongs to the first type of packet, the switch 120 may determine the output port number associated with the destination address of the packet and forward the first type of packet based on the output port number. message.

If the switch 120 determines that the packet to be forwarded belongs to the second type of packet, the switch may determine whether the packet is allowed to be forwarded to the destination device of the packet by performing an inbound processing operation on the packet. If the switch 120 determines that the message is allowed to be forwarded to the target device of the message via the switch, the switch 120 may forward the second output port number associated with the target address of the message based on the output port number. type message.

In order to further illustrate the simplification of the forwarding process of the first type message and the second type message, the difference between the simplified forwarding process and the conventional forwarding process will be described in detail below in conjunction with FIG. 5 . Fig. 5 shows a schematic diagram of a packet forwarding process according to an embodiment of the present disclosure.

As shown in Figure 5, the standard forwarding process 510 in DCN may include inbound port processing 511, inbound L2 processing 512, inbound access control list (ACL) check 513, L3 forwarding 514, forwarding associated data 515, outbound ACL check 516 , outbound direction L2 processing 517 and outbound port processing 518 .

In the forwarding process 520 for the second-type packet with the standard Ethernet frame format, the inbound port processing 521 can be omitted to reduce the delay. Optionally, only port attribute processing may be performed in the ingress port processing 521, so as to obtain all port attributes through a linear table (LT) lookup. In inbound L2 processing 522 , tri-state content addressable memory (TCAM) may be employed to perform inbound L2 processing. The matching specifications of the TCAM are small, which can further shorten the delay.

In some embodiments, the switch 120 may use exact match (EM) to find the output port number corresponding to the host address or network segment address during the L3 forwarding 523 process. In the same DCN, the address of the network segment is a configurable mask with a predetermined bit length.

In some embodiments, the switch 120 may also use the LT lookup to determine the corresponding output port number of the host address or network segment address during the L3 forwarding 523 process. Similarly, within the same DCN, the address of the network segment is a configurable mask with a predetermined bit length.

In this application, in order to implement a linear table lookup of IP addresses, a method of linearly arranging IP addresses may be used. For example, for a 32-bit IPv4 or 128-bit IPv6 address, only limited bits are used for IP address allocation, and in the forwarding process, only these bits are used for LT lookup, so as to achieve a faster lookup speed.

Afterwards, the switch 120 may perform the process of forwarding associated data 524 and the process of egress port processing 525 .

It can be seen that compared with the standard forwarding process 510, the forwarding process 520 for the second type message with the standard Ethernet frame format omits the inbound ACL check included in the standard forwarding process 510 in the DCN, and the outbound ACL Check and out direction L2 processing. Therefore, the complexity of packet processing during switch forwarding can be reduced, and the delay can be significantly reduced.

In the forwarding process 530 for the first-type packet with the simplified Ethernet frame format, the ingress port processing 531 can also be omitted to reduce the delay. Optionally, only port attribute processing may be performed in the ingress port processing 531, so as to obtain all port attributes through a linear table (LT) lookup.

Since the Ethernet header of the frame format of the first type of message includes the DMAC domain field with specific parameters, and no other L2 fields are included in the Ethernet header except the FCS field, the DCN field can be omitted in the forwarding process 530. The inbound direction L2 processing included in the standard forwarding process 510 .

L3 forwarding 532 in forwarding process 530 may perform similar operations as L3 forwarding 523 in forwarding process 520 . In some embodiments, the switch 120 may use exact match (EM) to find the output port number corresponding to the host address or network segment address during the L3 forwarding 532 process. In the same DCN, the address of the network segment is a configurable mask with a predetermined bit length. In some embodiments, the switch 120 may also use the LT lookup to determine the corresponding output port number of the host address or network segment address during the L3 forwarding 532 process. Similarly, within the same DCN, the address of the network segment is a configurable mask with a predetermined bit length.

Similarly, in order to implement a linear table lookup of IP addresses, a method of linearly arranging IP addresses can be used. For example, for a 32-bit IPv4 or 128-bit IPv6 address, only limited bits are used for IP address allocation, and in the forwarding process, only these bits are used for LT search to achieve a faster search speed.

Afterwards, the switch 120 may perform the process of forwarding associated data 533 and the process of egress port processing 534 .

Since compared with the forwarding process 520 for the second type message with the standard Ethernet frame format, the steps are further reduced in the forwarding process 530 for the first type message with the simplified Ethernet frame format, so in the message The delay caused by processing packets during forwarding can be further reduced.

According to the solutions of the embodiments of the present application, on the one hand, the NIC can implement a simplified Ethernet message without changing the original frame format by configuring parameters in specific fields in the frame format of the message. Therefore, the packets using the simplified Ethernet frame format and the packets using the standard Ethernet frame format can not only intercommunicate, but also coexist. This simplified Ethernet message can significantly reduce the complexity of the forwarding process, improve bandwidth efficiency and reduce delay.

On the other hand, the switch can determine the type of the message according to the parameter at a specific position in the message frame format and forward the message based on the determined message type. Therefore, the switch can implement different processing processes for different types of messages, for example, a simplified forwarding process can be used for simplified Ethernet messages, and other forwarding methods can be used for regular messages. In this way, accelerated forwarding of packets can be realized to reduce the time delay in the forwarding process.

Fig. 6 shows a flowchart of a method 600 for forwarding packets according to an embodiment of the present disclosure. The measurement method shown in FIG. 6 can be implemented at the switch 120 in FIG. 1 .

At block 610 , the switch 120 receives the message from the network interface card 110 .

At block 620, the switch 120 forwards the message according to configurable parameters in the destination MAC field in the frame format of the message. The configurable parameter is associated with the message type of the message, and the message type includes a first type message with a simplified Ethernet frame format and a second type message with a standard Ethernet frame format.

In some embodiments, the switch 120 may compare the configurable parameter with the predetermined parameter of the destination MAC field in the reference frame format of the first type packet. If it is determined that the configurable parameter matches the predetermined parameter, the switch 120 determines the received packet as the first type packet. Switch 120 may determine an output port number associated with the destination address of the message and forward the message based on the output port number.

In some embodiments, the switch 120 may compare the configurable parameter with the predetermined parameter of the destination MAC field in the reference frame format of the first type packet. If it is determined that the configurable parameter does not match the predetermined parameter, the switch 120 determines the received packet as the second type packet. The switch 120 may determine whether the packet is allowed to be forwarded to the target device of the packet by performing an inbound processing operation on the packet. If it is determined that the message is forwarded to the target device of the message via the switch, the switch 120 may determine an output port number associated with the target address of the message and forward the message based on the output port number. report message.

In some embodiments, the switch 120 obtains the target address from the frame format of the message, and the target address includes the address of the target device to receive the message, or the target network address related to the target device. segment address. The switch 120 determines the output port number by performing an exact match operation or a linear table lookup operation based on the target address.

In some embodiments, the address of the target network segment includes a configurable bitmask of predetermined length.

In some implementation manners, the switch 120 performs an inbound processing operation on the packet through a tri-state content-addressable memory matching operation.

By combining the method for forwarding packets described in FIG. 6 , the switch can implement different processing processes for different types of packets. In this way, accelerated forwarding of packets can be realized to reduce the time delay in the forwarding process.

Fig. 7 shows a flow chart of a method 700 for forwarding packets according to an embodiment of the present disclosure. The measurement method shown in FIG. 8 may be implemented at the NIC 110 in FIG. 1 .

At block 710, NIC 110 determines configurable parameters in the destination media access control field in the frame format of the message. The configurable parameter is associated with the message type of the message, and the message type includes a first type message with a simplified Ethernet frame format and a second type message with a standard Ethernet frame format.

In some embodiments, NIC 110 determines the type of data packet to be encapsulated into the message and determines a configurable parameter based on the type of data packet.

In some embodiments, if it is determined that the message to be generated is the first type message, the NIC 110 acquires a set of predetermined parameters of the destination MAC field in the reference frame format associated with the first type message And determining said configurable parameter from said set of predetermined parameters.

At block 720, the NIC 110 generates the message based on the configurable parameters.

At block 730, the NIC 110 sends the message to the switch.

By combining the method for message forwarding described in FIG. 7 , a simplified Ethernet message can be realized without changing the original frame format by configuring parameters in specific fields in the frame format of the message. Therefore, the packets using the simplified Ethernet frame format and the packets using the standard Ethernet frame format can not only intercommunicate, but also coexist. This simplified Ethernet message can significantly reduce the complexity of the forwarding process, improve bandwidth efficiency and reduce delay.

FIG. 8 shows a block diagram of a communication device 800 according to an example embodiment of the present disclosure. The communication device 800 may be implemented as the NIC 110 shown in FIG. 1 , the switch 120, a part of the above equipment or a chip in the above equipment, etc. It should be understood that the communication device 800 is used for exemplary purposes only and does not imply any limitation on the scope of the present disclosure. Embodiments of the present disclosure may also be embodied in different communication devices. In addition, it should also be understood that the communication device 800 may also include other elements, modules or entities, which are not shown for the purpose of clarity, but it does not mean that the embodiments of the present disclosure do not have these elements or entities. The scope of the present disclosure is not limited in this respect.

As shown in FIG. 8 , the communication device 800 includes an input and output interface 810 and a logic circuit 820 . The input-output interface 810 is coupled to a logic circuit 820 . In the embodiments of the present disclosure, the input and output interfaces 810 may be integrated together to implement the sending and receiving function, and may also be implemented as independent components as an input interface for receiving and an output interface for sending. For example, the input-output interface 810 shown in FIG. 8 is an example implementation of integration.

In an embodiment where the communication device 800 is implemented as a switch 120 as shown in FIG. 1 , the input and output interface 810 is configured to receive a message from the network interface card 110 and to Configurable parameters in forwarding packets. The configurable parameters are associated with the message type of the message, and the message type includes a first type message with a simplified Ethernet frame format and a second type message with a standard Ethernet frame format. In some example embodiments, the input-output interface 810 is further configured to compare the configurable parameter with the predetermined parameter of the destination MAC field in the reference frame format of the first type packet. If it is determined that the configurable parameter matches the predetermined parameter, the received message is determined as the first type message. An output port number associated with the destination address of the message is determined and the message is forwarded based on the output port number.

In some example embodiments, the input-output interface 810 is further configured to compare the configurable parameter with the predetermined parameter of the destination MAC field in the reference frame format of the first type packet. If it is determined that the configurable parameter does not match the predetermined parameter, then determine the received message as the second type of message. It is determined whether the packet is allowed to be forwarded to the target device of the packet by performing an inbound direction processing operation on the packet. If it is determined that the message is forwarded to the destination device of the message via the switch, determining an output port number associated with the destination address of the message and forwarding the message based on the output port number .

In some exemplary embodiments, the input and output interface 810 is further configured to obtain a target address from the frame format of the message, and the target address includes the address of the target device to receive the message, or is related to the The address of the target network segment related to the target device. The output port number is determined by performing an exact match operation or a linear table lookup operation based on the target address.

In some example embodiments, the address of the target network segment includes a configurable predetermined length bitmask.

In some exemplary embodiments, the I/O interface 810 is further configured to perform an inbound processing operation on the packet through a 3-state content-addressable memory matching operation.

In an embodiment in which the communication device 800 is implemented as the NIC 110 shown in FIG. 1 , the logic circuit 820 is configured to determine the configurable parameters in the destination MAC field in the frame format of the message and based on the configurable parameter The message is generated, wherein the configurable parameter is associated with a message type of the message, and the message type includes a first type message in a simplified Ethernet frame format and a second type message in a standard Ethernet frame format. . The input and output interface 810 may be configured to send packets to the switch.

In some example embodiments, the logic circuit 820 is further configured to determine a type of data packet to be encapsulated into the message and determine a configurable parameter based on the type of data packet.

In some example embodiments, the logic circuit 820 is further configured to obtain a set of predetermined parameters of the destination MAC field in the reference frame format associated with the first type of message and determine from the set of predetermined parameters The configurable parameters.

Figure 9 shows a block diagram of a communications device 900 in which certain embodiments of the present disclosure may be implemented. The communication device 900 can be used to implement the terminal device 110 and the network device 120 shown in FIG. 1 . The communication device 900 may also be implemented as a chip or a chip system. It should be understood that the communication device 900 is used for exemplary purposes only and does not imply any limitation on the scope of the present disclosure. Embodiments of the present disclosure may also be embodied in different devices. It should also be understood that the communication device 900 may also include other elements or entities, which are not shown for ease of description, but it does not mean that the embodiments of the present disclosure do not have these elements or entities.

As shown in FIG. 9 , the communication device 900 includes a processor 910 that controls the operations and functions of the communication device 900 . For example, in some example embodiments, processor 910 may perform various operations by means of instructions 930 stored in memory 920 coupled thereto. Memory 920 may be of any suitable type suitable for the local technical environment and may be implemented using any suitable data storage technology, including but not limited to semiconductor-based memory devices, magnetic memory devices and systems, optical memory devices and systems. Although only one memory unit is shown in FIG. 9 , there may be multiple physically distinct memory units in communication device 900 . It should be understood that the processor 910 and the memory 920 may be separately configured as individual components, or may be integrated together, which is not limited in this application.

Processor 910 may be of any suitable type suitable for the local technical environment, and may include, but is not limited to, general purpose computers, special purpose computers, microcontrollers, digital signal processors (Digital Signal Processor, DSP), and controller-based multi-core control One or more of the server architecture. Communications device 900 may also include multiple processors, such as application specific integrated circuit chips, that are time slaved to a clock that is synchronized with the main processor. The processor 910 is coupled with the communication unit 940 . The communication unit 940 may realize receiving and sending of information through radio signals or by means of optical fibers, cables and/or other components.

Memory 920 may include one or more non-volatile memories and one or more volatile memories. Examples of non-volatile memory include, but are not limited to, read-only memory (Read-Only Memory, ROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, EPROM), flash memory, hard disk, optical disc (Compact Disc, CD), Digital Video Disk (Digital Versatile Disc, DVD) or other magnetic and/or optical storage. Examples of volatile memory include but not limited to random access memory (Random Access Memory, RAM) and the like.

Embodiments of the present disclosure may be implemented by means of a computer program, enabling the communication device 900 to perform any of the processes as discussed with reference to FIGS. 2 to 7 . The embodiments of the present disclosure can also be realized by hardware or by a combination of software and hardware. The computer program comprises computer-executable instructions 930 executed by the processor 910 . Computer programs can be stored in the memory 920 . Processor 910 may perform any suitable actions and processes by loading a computer program into RAM.

In the embodiments described above where the NIC 110 or switch 120 determines tuning parameters by utilizing modules of machine learning algorithms, reinforcement learning models, etc., these modules and models may be stored in memory 920 in the form of computer program code or instructions 930 , the processor executes the program codes or instructions in the memory 920 to cause the communication device 900 to execute the processing procedures implemented by the NIC 110 or the switch 120 in FIG. 2 to FIG. 7 .

In some embodiments, the computer program may be tangibly embodied on a computer readable medium, which may be included in the communication apparatus 900 (such as in the memory 920 ) or other storage device accessible by the communication apparatus 900 . A computer program can be loaded from a computer readable medium into RAM for execution. The computer readable medium may include any type of tangible nonvolatile memory such as ROM, EPROM, flash memory, hard disk, CD, DVD, and the like.

In general, various example embodiments of the present disclosure may be implemented in hardware or special purpose circuits, software, logic, or any combination thereof. Certain aspects may be implemented in hardware, while other aspects may be implemented in firmware or software, which may be executed by a controller, microprocessor or other computing device. When aspects of example embodiments of the present disclosure are illustrated or described as block diagrams, flowcharts, or using some other graphical representation, it is to be understood that the blocks, devices, systems, techniques, or methods described herein may serve as non-limiting Examples of are implemented in hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controllers or other computing devices, or some combination thereof.

As an example, example embodiments of the present disclosure may be described in the context of machine or computer-executable instructions, such as program modules included in a device executed on a real or virtual processor of a target. Generally, program modules include routines, programs, libraries, objects, classes, components, data structures, etc. that perform particular tasks or implement particular abstract data structures. In various example embodiments, the functionality of the program modules may be combined or divided between the described program modules. Machine-executable instructions for program modules may be executed locally or in distributed devices. In a distributed device, program modules may be located in both local and remote storage media.

Computer program codes for implementing the methods of the present disclosure may be written in one or more programming languages. These computer program codes can be provided to processors of general-purpose computers, special-purpose computers, or other programmable data processing devices, so that when the program codes are executed by the computer or other programmable data processing devices, The functions/operations specified in are implemented. The program code may execute entirely on the computer, partly on the computer, as a stand-alone software package, partly on the computer and partly on a remote computer or entirely on the remote computer or server.

In the context of the present disclosure, computer program code or related data may be carried by any suitable carrier to enable a device, apparatus or processor to perform the various processes and operations described above. Examples of carriers include signals, computer readable media, and the like. Examples of signals may include electrical, optical, radio, sound, or other forms of propagated signals, such as carrier waves, infrared signals, and the like.

In the context of the present disclosure, a machine-readable medium or computer-readable medium may be any tangible medium containing or storing a program for or relating to an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination thereof. More detailed examples of machine-readable storage media include electrical connections with one or more wires, portable computer diskettes, hard disks, random storage access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash), optical storage, magnetic storage, or any suitable combination thereof.

In addition, while operations are depicted in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown, or in sequential order, or that all illustrated operations be performed, to achieve desirable results. In some cases, multitasking or parallel processing can be beneficial. Likewise, while the above discussion contains certain specific implementation details, these should not be construed as limitations on the scope of any invention or claims, but rather as a description of specific example embodiments that may be directed to particular inventions. Certain features that are described in this specification in the context of separate example embodiments can also be implemented integrally in a single example embodiment. Conversely, various features that are described in the context of a single example embodiment can also be implemented in multiple example embodiments separately or in any suitable subcombination.

Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.

Claims

A method for forwarding a message, comprising:

The switch receives the message from the network interface card;

The switch forwards the message according to a configurable parameter in the destination media access control field in the frame format of the message, where the configurable parameter is associated with the message type of the message, and the The message types include a first type message with a simplified Ethernet frame format and a second type message with a standard Ethernet frame format.
The method according to claim 1, wherein forwarding the message comprises:

comparing said configurable parameters with predetermined parameters of a destination media access control field in a reference frame format of said first type of message;

If it is determined that the configurable parameter matches the predetermined parameter, determine the received message as a first type message;

determining an output port number associated with the destination address of the message; and

Forwarding the message based on the output port number.
The method according to claim 1, wherein forwarding the message comprises:

comparing said configurable parameters with predetermined parameters of a destination media access control field in a reference frame format of said first type of message;

If it is determined that the configurable parameter does not match the predetermined parameter, determining the received message as a second type message;

determining whether the packet is allowed to be forwarded to the target device of the packet via the switch by performing an inbound processing operation on the packet;

If it is determined that the message is allowed to be forwarded to the destination device of the message via the switch, determining an output port number associated with the destination address of the message; and

Forwarding the message based on the output port number.
The method according to claim 2 or 3, wherein determining the output port number comprises:

Obtaining a target address from the frame format of the message, where the target address includes one of the following:

the address of the target device to which the message is to be received, or

the address of the target network segment associated with the target device; and

The output port number is determined by performing an exact match operation or a linear table lookup operation based on the target address.
3. The method according to claim 2 or 3, wherein the destination address of the message is an Internet Protocol address having a predetermined number of bits.
The method according to claim 4, wherein the address of the target network segment comprises a configurable predetermined length bitmask.
The method according to claim 3, further comprising:

An inbound processing operation is performed on the message through a tri-state content addressable memory matching operation.
A method for forwarding a message, comprising:

A network interface card determines a configurable parameter in a destination media access control field in a frame format of a message, wherein the configurable parameter is associated with a message type of the message, the message type including a The first type of message in the Ethernet frame format and the second type of message in the standard Ethernet frame format;

the network interface card generates the message based on the configurable parameters; and

The network interface card sends the packet to the switch.
The method of claim 8, wherein determining a configurable parameter comprises:

determining the type of data packet to be encapsulated into the message; and

The configurable parameter is determined based on the type of the data packet.
The method of claim 8, wherein determining the configurable parameter comprises:

Obtaining a set of predetermined parameters of the destination MAC field in the reference frame format associated with the first type of message; and

The configurable parameter is determined from the set of predetermined parameters.
A switch, comprising:

processor; and

A memory coupled to the processor, the memory stores instructions that need to be executed, and when the instructions are executed by the processor, the switch executes the method according to any one of claims 1-7.
A network interface card comprising:

processor; and

A memory coupled to the processor, the memory stores instructions that need to be executed, and when the instructions are executed by the processor, the network interface card executes the method according to any one of claims 8-10.
A computer readable storage medium. The computer-readable storage medium stores instructions, which when executed by the electronic device cause the electronic device to perform the method according to any one of claims 1-7 or the method according to any one of claims 8-10.
A computer program product. The computer program product comprises instructions which, when executed by an electronic device, cause the electronic device to perform the method according to any one of claims 1-7 or the method according to any one of claims 8-10.